Decreasing the Assimilation Effect of the Delboeuf Illusion

Craig Skinner
00caskin@alma.edu
Alma College

Bethany Simmons
00besimm@alma.edu
Alma College

The Delboeuf illusion consists of a perceived difference in size between two shapes of equal size when one of the shapes is accompanied by a larger concentric shape, and the other shape contains a smaller concentric shape. It is widely agreed upon that this is due to the assimilation effect (Weintraub, Wilson, Greene, & Palmquist 1969).

This experiment further tests assimilation as the explanation of the Delboeuf Illusion. This experiment is also intended to show where a crossover takes place from assimilation to contrast. We hypothesized that as the medium-sized figure (T) is displaced farther from the center-point of the large figure (L), the assimilation effect would decrease. Also, when it is completely outside and no longer touching, a contrast effect (reversal of the usual illusion) would occur.

Methods

51 students, ages 18-22, participated in this experiment which was conducted in the Eyelines computer program (Beagley 1990). Subjects were seated with their eyes a distance of 1 and 1/2 feet from the screen.

The stimuli were eight different versions of the Delboeuf illusion (Figure 1). T consisted of lines of 60 mm; L consisted of lines of 90 mm; and S consisted of lines of 40 mm.

Subjects were presented with each stimuli twice in random order in which C varied up to one-fifth of the size of T. The subjects were asked to manipulate C so that it would be the same apparent size as T.

The mean illusion error was calculated for each trial. Control's error was subtracted from Contrast Control's error. This number was subtracted from the error of each variation of the Delboeuf illusion (successive orientations of T in relation to the center point of L).

Results

As hypothesized, the farther T was moved away from the center of L, the weaker the illusion became.

Control and Contrast Control resulted in subjects undersestimating the size of T and therefore making C smaller than it should have been. The standard Delboeuf Illusion and manipulated versions of it resulted in subjects overestimating the size of T, thus making C larger than it should have been.

Figure 3 shows the mean illusion magnitude of each trial which was controlled for by the Contrast Control (Control was first subtracted from Contrast Control). This controlled for the possibility of a contrast effect which proved not to be evident in this experiment. Clearly, the strongest illusion was produced by the standard Delboef Illusion (Concentric). As T was moved farther away from the center of L, each subsequent trial resulted in decreased illusion magnitude.

Discussion

Results suggest that the assimilaiton effect is decreased by shifting T farther from the center point of L, but at no point does the effect completely disappear. Furthermore, at no point does the hypothesized contrast effect occur. The results suggest that two shapes do not have to be concentric for our visual system to assimilate them. Assimilation also may take place when two shapes are merely next to each other. It is also possible that there is another factor influencing the illusion. Studies have shown that reducing the reflectance contrast of T increases the illusion (Weintraub, Wilson, Greene, Palmquist 1969). Results such as these suggest an interaction effect between lateral inhibition and assimilation.

References

Beagley, W. K. Eyelines [Computer Program]. Alma, MI. Alma College.

Weintraub, D.J., Wilson, B.A., Greene, R.D., and Palmquist, M.J. Delboeuf Illusion: Displacement versus diameter, arc deletions and brightness contrast. Journal of Experimental psychology, 1969, 80, 3, 505-511.